Critical Review of Text Mining and Sentiment Analysis for Stock Market Prediction

Abstract

The paper is aimed at a critical review of the literature dealing with text mining and sentiment analysis for stock market prediction. The aim of this work is to create a critical review of the literature, especially with regard to the latest findings of research articles in the selected topic strictly focused on stock markets represented by stock indices or stock titles. This requires examining and critically analyzing the methods used in the analysis of sentiment from textual data, with special regard to the possibility of generalization and transferability of research results. For this reason, an analytical approach is also used in working with the literature and a critical approach in its organization, especially for completeness, coherence, and consistency. Based on the selected criteria, 260 articles corresponding to the subject area are selected from the world databases of Web of Science and Scopus. These studies are graphically captured through bibliometric analysis. Subsequently, the selection of articles was narrowed to 49. The outputs are synthesized and the main findings and limits of the current state of research are highlighted with possible future directions of subsequent research.

Temporal Convolutional Networks and BERT-Based Multi-Label Emotion Analysis for Financial Forecasting

“…Moving on to papers regarding financial forecasting and sentiment analysis, a comprehensive literature-based study on investor sentiment analytics and machine learning applied to predict stock prices is presented in [19]. Additionally, review-wise, the work in [20] presents a critical literature review regarding text mining and sentiment analysis for stock market prediction, focusing on stock markets. A systematic review examining works based on using machine learning and text mining techniques applied to news data to predict the stock market is presented in [21].…”